Question: I am into storage administration and learning about the SAN and NAS environments. There is much talk about SATA storage taking over SCSI. How does SATA differ in its functionality with SCSI?Randy Kerns' answer:
There are many differences in the interface technology and in the commands executed. SCSI has a robust tagged command queuing implementation to allow multiple commands to be outstanding which provides significant performance gains for drives or controllers by being able to order the commands in an optimal execution manner. ATA only allows a bus release that is mistakenly called command queuing. The implementation on the class of disk drives that support ATA is different as well.
ATA drives are cheaper than SCSI or Fibre Channel drives and there's a reason. SCSI and FC drives use a processor for executing the commands and handling the interface and a separate processor controlling the head positioning through servos. ATA drives use a single processor for both which means that if the rotational positioning requires more adjustments due to factors such as rotational vibration or wear, more processor time will be dedicated for that which can affect the performance of the drive. Because of this, you will usually see a lower RPM for the ATA disk drive and also a shorter warranty period. But it is cheaper and can be a very good, economical solution for many environments and applications.
The "take over" term you used is probably some marketing BS or from someone who doesn't understand fully. There are justifications for both higher performance and less expensive disks. The real issue is the software to manage placement of data on the appropriate devices.
Ed note: If you would like to read additional compliance articles, opinions and expert advice, make sure to sign-up for our ALERTS on compliance. Click here to sign up. SearchStorage.com also offers alerts on low-cost storage.
This was first published in February 2004